1. Field of the Disclosure
The present disclosure relates to the field of data center virtualization and, more particularly, to systems and methods for providing dynamic operational integrity attestation of application security and a user reputation at runtime.
2. Description of the Related Art
One recent trend in computing is the trend towards virtualization and cloud computing in which, for example, enterprise software is no longer owned by the customer, but instead the enterprise's Information Technology (IT) infrastructure can be provided by a third party and enterprise software applications are sold as service offerings.
Traditional legacy and currently available security technologies, such as next-generation anti-virus software, network firewalls, and intrusion detection/prevention systems, operate based on signatures, protocol anomalies or virtual execution in a sandbox to monitor threats and attacks. However, once a target of an attack or threat (i.e., a ‘target system’ or ‘target platform’) becomes infected or compromised, many of these technologies have been proven relatively ineffective in protecting systems and preventing data breaches or service disruptions. Emerging threats and attacks can exhibit a low and slow mode of operation and can be signature-less in order to evade traditional detection and defense technologies. Further, these technologies are often agnostic to the runtime operational integrity of computer workloads on the target (victim) systems and therefore often do not offer any level of remediation to service the affected target systems. As such, many traditional security technologies merely provide coarse-grained access controls limited to a simple allow or deny decision logic.
Many enterprises and organizations are moving there IT infrastructures to cloud computing environments (from self-managed on-premise data centers to service-provider managed outsourced virtual data centers), wherein which third parties may provide shared computing resources and applications running on those resources. They are being offered as services to a plurality of customers. The move by companies to cloud computing environments are increasing for various reasons, including the need for increased computing power, increased storage capacity, and/or increased network bandwidth, among others. Enterprise applications and mission critical applications may be executed in the cloud. Without adequate device, system, and application security, the cloud can compromise these applications, potentially causing large financial losses. Data confidentiality, data integrity and data availability can be maintained in the cloud computing environment even when such applications may be controlled by these third parties. The cloud may enable traditional information technology devices, systems and applications to transparently execute from these service-providers managed outsourced virtual data centers. As newer technologies emerge, the cloud infrastructure may evolve transparent to and scalable with enterprise operations. Security in such environments may be an emerging challenge. For example, virtualized, cloud computing on-demand, and elastic models require dynamic security orchestration at both the device and network level that is lacking in traditional security approaches and technologies.
A significant metric that relates to the lack of remediation controls in current security controls is the high rate of false positives and negatives in the detection of threats. False positives are unproductive and reduce operating capacity in data centers. False negatives lead to the bypass of security controls and compromise of the target systems and services. Existing signature-based approaches to security control are vulnerable to improvised attacks staged at targeted systems.
The proliferation of applications (business, social networking and gaming software) in Enterprise and cloud computing ecosystems has significantly increased the attack surface and window of exposure. Application runtime operational integrity is a critical factor in building end-to-end trust from device to service. The proliferation of unmanaged devices such as mobile computing devices (i.e., tablets and smart phones) and Bring Your Own Device (BYOD) policies in Enterprise network environments has increased the risk of advanced coordinated threats.
The current predominant security paradigm is based on a hard edge and soft core architecture. Security appliances such as network firewalls and intrusion detection/prevention systems deploy at the edge (perimeter). Antivirus and network based integrity measurement and verification services scan and audit the soft-core that comprises business critical systems, services and high value data silos. Once the edge is breached, defensive methods are largely ineffective in protecting vulnerable or compromised systems. In this paradigm, the hard edge does not generally provide any level of assurance of the runtime operational integrity of the soft core.
The heavy reliance on extensive application whitelists/blacklists(based on file hash digests and/or configuration), Internet Protocol (IP) address reputation lists, signatures, protocol anomalies and virtual execution in a sandbox, provides targeted and coordinated attacks a large staging surface to maneuver around with meticulously engineered evasion methods exploiting the gaps. Irrespective of the method of break-in, the post-infection behaviors during the window of exposure on a victim machine or environment are difficult to conceal, but may be obscured from timely detection and diagnosis by security administrators and virtual execution (sandbox) environments through evasion techniques. Various exemplary embodiments include an early warning system for threat identification and diagnosis with high forensic confidence and infection summaries for manual and/or automated intervention.
Current identity management (IdM) approaches are typically aimed at domain level authentication of users for the use of a security token in a transaction or business process wherein the issuance of the token is strictly based on proof of possession of credentials and not the reputation of the user. While the industry has adopted multi-factor authentication technologies for increased protection from malicious users whose intent is to steal credentials for the purpose of impersonation or delegation, these security controls do not provide a means for incorporating attribution of an authenticated user's risk posture, independent of provisioned entitlements (i.e., user account privileges and roles). Thus, existing security controls do not incorporate a complete view of a user's risk posture as a component of an access policy, based on correlation of location and device agnostic global threat intelligence about the user. Accordingly, what is needed are systems and methods for runtime operational integrity monitoring of applications by providing dynamic operational integrity attestation of application security and user reputation at runtime that take into account a subject's (such as a device, application, or user) risk posture. What is further needed are systems and methods for performing post infection diagnoses, threat identification via signature-less anomalous behavior recognition using non-intrusive platform instrumentation without requiring ‘hooks’ into an operating system (OS) Kernel by security vendors that leverages OS vendor application programming interfaces (APIs).
The emerging cloud based application-hosting model requires a higher level of scrutiny of device and user security posture. Threats can be posed by both external and internal users, and therefore it has become harder to determine whether an authenticated user is a trustworthy operative. Application and network layer entitlements are largely static in nature. While role based access control (RBAC) systems are positioned to offer dynamic policies, these are not based on any form of aggregation of external threat intelligence in the context of the user in the transaction.
Tokenization of identities by Security Token Services (STS) intended to facilitate in Identity Federation and Single Sign On (SSO) for web and Enterprise applications have only aggravated the threat vectors by expanding the scope of resources that an authenticated user may access implicitly without granular context. Along with the coarse access policies offered by traditional security controls, such as firewalls and intrusion prevention systems, this increases the attack surface for application exploits by malicious users leveraging the excessive privileges granted by default to most users (and groups) today, in the absence of strict Separation of Duties (SoD) and Principle of Least Privilege (PoLP) enforcement.
Several studies have uncovered conclusive evidence that the largest risk, associated with malicious activities resulting in theft of confidential data, intellectual property or financial losses for enterprises and consumers, arises from insider rather than outsider or man-in-the-middle attacks. Threat assessment systems have been predominantly focused on the means and methods rather than the adversary or operator (i.e., a user). There has been no emphasis on the reputation score of a user in the transaction or business process as a real time integrity metric for integration with access policy decision logic. Accordingly, what is further needed are systems and methods that employ a user reputation score that can serve either as a punitive or corrective method of intervention for the proactive prevention of exploits and data breaches.
One goal of trusted computing is to provide a resource owner or service provider with reliable knowledge about a target system. Current attestation systems are components of computer systems that permit reliable statements of evidence about those systems to be conveyed to remote parties, including other computers. Through evaluation of the identity and integrity of components of a system (i.e., a target system being evaluated), evidence is produced that the target system will not engage in defined classes of misbehaviors. As the users accessing a system and the requirements of applications running on a system generally cannot be known a priori, attestation systems and measurement systems alike must be flexible, providing for privacy, completeness of measurement, and trust in the basic collection and reporting mechanisms.
Existing attestation systems are often narrowly focused and generally aimed at specific use-cases limited to system components such as hardware and applications and therefore typically lack flexibility to dynamically address more general attestation problems. Further, existing definitions of attestation focus primarily on describing specific, narrow, and particular properties desirable in specific use-cases. Additionally, current attestation systems are created to work with one particular measurement system targeting one particular system of interest without considering the reputations of users or operators of the system of interest.
Accordingly, what the present inventors have identified as desirable are technology based attestation architectures, systems, and methods that perform a calculus of risk to determine a level of trust of systems that are not necessarily monolithic and can be made up of diverse hardware and software platforms and can be accessed by users with varying credentials and reputations. Accordingly, what the inventors have identified as being desirable are attestation architectures, systems, and methods that can be flexible enough to accommodate varying concepts of attestation, including taking the reputation of users that access targeted systems into account. What the present inventors also see as desirable are attestation systems and architectures that can dynamically handle complex attestation scenarios and provide more complete runtime attestation than is currently achievable. What the present inventors additionally see as desirable are methods and systems for performing evidence based automated application binary analysis.